1. Post-processing Overview Bryan Butler, for Joe McMullin

2. Bryan ButlerEVLA Advisory Committee Meeting May 8-9, 20062 Science Software Group Mission: Primary purpose is to provide facilities for post- observational scientific reduction of ALMA and EVLA data Facilities must be: Scientifically complete/correct Robust Acceptably fast Easy to use Able to support both scripted and interactive use

3. Bryan ButlerEVLA Advisory Committee Meeting May 8-9, 20063 SSG - Requirements Requirements Key Areas: General Requirements and Interaction Interface Data Handling Calibration and Editing Imaging Data Analysis Visualization Special Features Driven by requirements documents: –ALMA Offline Data Processing Requirements –EVLA Data Post-processing Software Requirements –2/3 of requirements are common! Detailed tracking on the combined projects’ requirements (common to both projects and the deltas between projects)

4. Bryan ButlerEVLA Advisory Committee Meeting May 8-9, 20064 SSG - Organization The SSG is a group within the Interferometry Software Division (ISD) - key personnel for ISD and SSG are: G. van Moorsel (EVLA), B. Glendenning (ALMA) - Heads of Computing; ISD Co-managers M. Rupen (EVLA), Debra Shepherd (ALMA) - Project Scientists for Computing J. McMullin, K. Golap - Group Management S. Myers - Group Project Scientist ISD functions now in the new e2e Operations Division: N. Radziwill, E. Fomalont - manager and scientist

5. Bryan ButlerEVLA Advisory Committee Meeting May 8-9, 20065 SSG - Staffing NRAO : 5.0 FTEs (including management; split 60/40 between ALMA & EVLA) EVLA-specific: 1.5 FTEs (Imaging algorithms, Data flagging) ALMA-specific : 5.5 FTEs (Application development, Data Model, Data Model Interfaces, Data Capture process) NSF grant : 1.0 FTE (visualization; concludes 2007.8) More detail on backup slide, if needed

6. Bryan ButlerEVLA Advisory Committee Meeting May 8-9, 20066 SSG - Software Packages Packages: Legacy AIPS++ code base for development (changing to CASA - more on that later) AIPS used for validation, algorithm and display development (e.g., automated flagging), support of EVLA hardware tests

7. Bryan ButlerEVLA Advisory Committee Meeting May 8-9, 20067 AIPS++ - Transition to CASA Users reluctant to embrace AIPS++ as a post-processing package, based on problems with robustness, speed, functionality, and the user interface Decision to halt public releases of AIPS++ made in 2004 SSG now concentrating on above problems; also working on modularity of code, a change from Glish to Python as the binding scripting language, and help/documentation New package named CASA - Common Astronomical Software Applications Care being taken regarding releasing the new package openly, to avoid the problems of the past

8. Bryan ButlerEVLA Advisory Committee Meeting May 8-9, 20068 Framework Improvement AIPS++: Glish interface; relatively unknown, unsupported outside of NRAO Tasking system based on Glish GUI system based on Glish/Tk; limited widgets, not robust! Difficult for external developers to contribute Multi-CD binary distribution Large monolithic libraries with cross dependencies Freeze 2 nd half of 2006 CASA : Hierarchical set of small libraries with clearly defined dependencies Inherits all application code improvements in robustness and performance. Smaller memory footprint/startup time Python interface (community standard); IPython Binding to Python, ACS; other frameworks enabled

9. Bryan ButlerEVLA Advisory Committee Meeting May 8-9, 20069 CASA Usage Status Has been tested internally by NRAO scientists in preparation for the 2006 ALMA test. –Early demos provided bi-monthly to the NAUG at: http://casa.nrao.edu/gettingstarted.shtml –6 scientists used and reported on it – enabled deployment for the ALMA test. Was deployed and reviewed by four ALMA testers –http://projectoffice.aips2.nrao.edu/ALMA2006.01/ALMA2006.01.html The user interface was reviewed and commented on by 8 NRAO scientists –User Interface is being refined/further developed based on the user interface report: http://projectoffice.aips2.nrao.edu/uiwg-report.pdf Will be used exclusively by the ALMA Pipeline Heuristics Team (second half of 2006) CASA will have replaced AIPS++ within NRAO (developers and NAUG testers) this year.

10. Bryan ButlerEVLA Advisory Committee Meeting May 8-9, 200610 CASA Development/Release Plan 2006 –NRAO: NAUG testing (AIPS++) –Community: Project tests –ALMA: external science testing: +single-baseline commissioning test) (CASA) –ALMA: external science testing: +single dish reduction (CASA) –ALMA: commissioning support (CASA) –ALMA/EVLA: user interface review (CASA) –EVLA: external science testing: +full polarization imaging; antenna pointing calibration –2006.5 AIPS++ frozen –ALMA: Pipeline Heuristics Use of CASA –NRAO: CASA initial beta deployment (internal) 2007 –NRAO: User support (CASA) –ALMA: commissiong support (CASA) –EVLA: external science testing: + RFI/automated flagging (CASA) –ALMA: P1 SSRs complete –NRAO: CASA second beta deployment (internal) 2008 –EVLA: external science testing: + wide band calibration imaging (CASA) –ALMA: commissioning support –EVLA: P1 SSRs complete –Community: CASA mature - limited release to outside users 2009-2010 –ALMA/EVLA: commissioning support –Community: CASA released/distributed for early ALMA/EVLA science 2011 –Community: CASA released/distributed; full user support (ALMA/EVLA: P1/P2s)

11. Bryan ButlerEVLA Advisory Committee Meeting May 8-9, 200611 SSG - Ability to Deliver The SSG has demonstrated the ability to meet major milestones: Heavily reviewed in annual VC and UC meetings, as well as ALMA CIPT, ASAC and EAC reviews Of 21 major milestones presented at ALMA CDR2 in July 2004 for the period up until now, 15 were met on time, 2 were changed and met on time, and 4 were met, but late Over past three years, 3 formal ALMA CDRs, 4 formal ALMA releases, 3 formal ALMA tests, and 1 formal EVLA test have been successfully completed Note that EVLA benefits substantially from ALMA-driven development of CASA, since common requirements are met (remember that 2/3 are common), as well as the higher-level requirements of robustness, speed, and usability.

12. Bryan ButlerEVLA Advisory Committee Meeting May 8-9, 200612 User Interface Significant effort has been devoted to the developing the command line version of the UI. It was tested extensively in early April 2006, with the following results: Prototype IPython interface to functionality, parameter setting and in- line help are in the right direction but more development and user testing needed. Full needs are documented at: http://projectoffice.aips2.nrao.edu/uiwg-report.pdf Other important (high priority) user interface issues: A revised interface prototype (similar to IRAF epar environment) is being developed There is much work needed to provide astronomer-level documentation (collaboration with NAUG to develop this) Development progress will be reviewed with the Fall EVLA test, as well as less formal, on-going testing.

13. Bryan ButlerEVLA Advisory Committee Meeting May 8-9, 200613 Risks Any reduction in staffing levels within the SSG can impact delivery schedule –Staff retention is key; AIPS can potentially be used to bridge any delivery schedule gap The necessity to do both software development and active user support stretch already thin resources –Convert to NRAO staff scientists being front-line user support, just as for AIPS User interfaces are notoriously difficult to develop (and gain acceptance) –Extensive testing and feedback into development is critical Significant scientific staff resources are required for testing, support, and documentation; not present currently –NRAO intends to redirect or hire new scientific staff to address this Science Data Model complexity can impact delivery schedule –The SDM must be critically evaluated for propriety and agreed to by both ALMA and EVLA (and potentially GBT) Algorithm development and data rates and computation –Algorithm development group (NAWG) work ongoing; data rate and computation problems addressed with additional hardware

14. Bryan ButlerEVLA Advisory Committee Meeting May 8-9, 200614 Resources - Details FTEsComments NRAO (SSG)4.75 [5.75]Project Scientist: Steve Myers Sanjay Bhatnagar0.75AstronomerAlgorithm Development (EVLA) Kumar Golap0.75AstronomerApplication CodeApplication code: common to ALMA&EVLA Joe McMullin0.75AstronomerManagement George Moellenbrock0.75AstronomerApplication Code David DeBonis1.0Software EngineerSystemBegan 2004-05; depart:2006-01 Darrell Schiebel1.0Software EngineerSystem Wes Young0.75Software EngineerSystem ALMA4.3 [5.3]2.3 FTE in-kind Gary Li1.0Software EngineerApplication Code0.5 FTE in-kind; Began 2004-03 Raymond Rusk1.0AstronomerApplication Code Tak Tsutsumi1.0AstronomerApplication Code1.0 FTE in-kind; Began 2005-03 (ACA) Honglin Ye0.5Software EngineerApplication CodeBegan 2005-07 (visualization) Michel Caillat0.8Software EngineerALMA Data model, simulation0.3 FTE in-kind Francois Viallefond0.5AstronomerALMA Data model, simulation0.5 FTE in-kind Heiko Hafok0.5AstronomerALMA Data CaptureBegan 2003-12 NSF grant1.0 David King1.0Software Engineervisualization(concludes 2006-05) Other0.5 Urvashi Rao0.5AstronomerAlgorithm Development (EVLA)Began 2004-09 (9 month appt) Total10.05 [12.55]